<mollyboo:> There is no trick to pawn storms. It's the slowest possible attack, and the only way it wins is if your opponent is trying an even slower one on the other side of the board. *boomp-chh*

Some openings that regularly feature pawn storms are:

- King's Indian (before 2006?), Black storms the Kingside (note how modern Bayonet Attack is a White counter-storm on the Queenside, targeting d6 with heavy-invasion at c7)

- Sicilian with opposite castling, White storms the Kingside, Black storms the Queenside

- Sicilian, Dragon, Yugoslav Attack (before 1990?), White mini-storms the Kingside (and did well enough to prompt Fischer's boast of "sac, sac, mate")

Play through lots of those games to get a feel for the pawn storms. The date limits indicate advances in modern theory that has caused some lines to fall out of favor -- e.g. Black's KID pawn storm has scored so well in the past that nowadays everybody plays anti-storm lines.

(Here's a grand puzzle for all chessgames.com readers: On which positional clues should Black <not> try the KID pawn storm, and switch to Q-side Benkish pressure instead?)

The uber-theme behind all pawn storms is your piece strength behind the pawns. Pawns do nothing by themselves. Your goal is always to open one line, and then exploit with your pieces.

Cases in point:

A. A couple of canonical sac patterns involve advanced pawn + piece sac: <Greco> Kh8, Rh1, h5, Nh4: Ng6(+) to open h, <swinging gate> Kh8, Rg1, Rh1, h6: Qg7+!! to open h with double-check and a triangle mate: Capablanca vs T A Carter, 1909, Stanishevsky vs Nikonov, 1981. Mnemonic: <Push that P to 6 or 3!>

B. Yugoslav Attack lines (before everybody and his whipped puppy learned not to let White do this): The point of White's h5-hxg6 is to half-open h for Qh6+: <It's all about the Queen entry.> Be3-Qd2-Bh6 is preparation, Bc4-Bb3 pins f7, and if White gets half-open h, then Black has a <Qh8 dilemma>. This is a classic example of pieces and a pawn storm all working harmoniously. Note how White spends about 6 piece moves to prepare the threat, then pushes with just one pawn, and follows up with piece strength.

White to play (25.?) "Insane" (4 stars)

It's nice to see a mini-tribute to Aleks Wojtkiewicz this week; he died much too young.

The only thing that seems worth analyzing is 25. Rd7! Bxd7 (pretty forcing since 25...g6 is answered well by 26. Qxg6+ fxg6 27. Rxc7 with a healthy pawn up and rook on the 7th for a big endgame advantage) 26. Qxf7+ and now:

a) 26...Kh7 27. Nxd7, threatening 28. Qxe8 as well as 28. Nxf8+ winning the queen; if 27...Rc8 28. Be5! (even stronger than Nxf8+) Bd6 (or 28...Qd8 29. Nf6+ gxf6 30. Bxf6+) 29. Nf6+ Kh8 30. Qg6 gxf6 31. Bxf6+ Ng7 (Qg7 lasts longer, but is still losing) 32. Be4 Kg8 33. Qh7+ Kf8 34. Qh8+ Kf7 35. Qxg7+ Ke8 36. Bg6+ and mate next move; if 27...Qc8 28. Be4! and there is no defense to 29. Nxf8+ and 30. Qxg7#.

I instantly saw 25.Rd7 because black can't capture it with the Queen, so Bishop comes in between(Capturing the Rook). Then the Pawn on f7 is undefended. Now comes 26.Qxf7 and black King goes to h8.

Here I went with 27.Ng6+ and thought that after 27...Kh7 comes 28.Nxf8 with Check and I assumed that black would not capture the knight with his rook and I would be generating windmills. But why wouldn't he(Gender Neutral)? After 28...Rxf8 29.Qxf8 gxf6 Black is a piece up.

Then I saw the solution. I realized that 27.Be4 was the move instead of 27.Ng6+ and I also realized that after 27.Ng6+ Kh7 28.Nxf8 black simply plays 28...Rxf8

I knew that this 28.Nxf8 move was silly but I still ran it through Stockfish for the peace of my mind. And Guess what, I found two new drawing variations which are given below -

a) If Black takes the Bishop - 29.Qxf8 gxf6 30.Qxf6 Bc8 31.Be4 Qxc4 32.Qf7+ Kh8 33.Bxf5 Qc1+ 34.Kg2 exf5 35.Qf8+ Kh7 36.Qf7+ Kh8 =0.00 (29 ply) b) If Black doesn't take the Bishop - 29.Qxf8 Qxc4 30.h3 Qc1+ 31.Kh2 Qd2 32.g4 Qf4+ 33.Kh1 Qc1+ 34.Kh2 Qf4+ =0.00 (28 ply)

The first idea that comes to mind is 25.Rd7 Bxd7 26.Qxf7+:

A) 26... Kh7 27.Nxd7

A.1) 27... Qc8 28.Bxb7

A.1.a) 28... Qxb7 29.Nxf8+ and 30.Qxb7 wins decisive material.

A.1.b) 28... Re7 29.Qg6+ and 30.Bxc8 wins decisive material.

A.2) 27... Rc8 28.Nxf8+ Kh8 (28... Rxf8 29.Qxc7 wins) 29.Ng6+ Kh7 30.Ne7 looks winning (30... Nxe7 31.Qxg7#; 30... Rb8 31.Qg6+ Kh8 32.Bxg7+ Nxg7 33.Qxh6#).

B) 26... Kh8 27.Be4 (27.Ng6+ Kh7 28.Be4 Qd6 and Black seems to hold)

B.1) 27... Qd6 28.Nxd7

B.1.a) 28... gxf6 29.Qxe8 Kg7 30.Qxf8+ Qxf8 31.Nxf8 Kxf8 32.Bxb7 wins decisive material.

B.1.b) 28... Rc8 29.c5 Qd1+ 30.Kg2

B.1.b.i) 30... gxf6 31.Nxf8 Ng7 32.Ng6+ Kh7 33.Ne7+ Kh8 34.Qg6 wins.

B.1.b.ii) 30... Ne3+ 31.fxe3 Qe2+ 32.Kh3 looks winning for White.

B.2) 27... gxf6 28.Ng6#.

click for larger view

It seems like if 30.Qg6 Nxe4 31.Qxe8 Nxf6 32.Nxf6 Qd6 occurs, it is not that bad.

click for larger view

It was fairly common when they would show him their games that he would for example say immediately after a move, something like: "there is mate in 5" or "black wins material in 4". He would show this often after he had for an hour or so been drinking the student's liquor too!

It seems like if 30.Qg6 Nxe4 31.Qxe8 Nxf6 32.Nxf6 Qd6 occurs, it is not that bad.>

1) -1.25 (27 ply) 28...exf5 29.Nxd7 Qe6 30.Qxe6 Rxe6 31.Be5 Re8 32.f4 Kg8 33.Kf2 Rd8 34.Nxf8 Kxf8 35.Ke3 g5 36.a4 Kf7 37.c5 Ke6 38.b5 axb5 39.axb5 Rd5 40.Bd6 gxf4+ 41.Kxf4 Rd2 42.h3 Rd4+ 43.Ke3 Kd5 44.Kf3 Rc4 45.b6 Kc6 46.Ke3

The text move Nxd7 is winning, although the computer actually likes Kg2 also.

After 28 Nxd7:

1) +3.91 (24 ply) 28...Qd1+ 29.Kg2 Nd6 30.Qg6 Nxe4 31.Qxe8 Nxf6 32.Nxf6 Qd6 33.Nd7 Kg8 34.c5 Qd5+ 35.f3 Qf5 36.g4 Qc2+ 37.Kh3 Qe2 38.Qxf8+ Kh7 39.Nf6+ gxf6 40.Qf7+ Kh8 41.Qxf6+ Kg8 42.Qg6+ Kf8 43.Qxh6+ Kf7 44.f4 Qc2 45.Kg3 a5 46.Qh5+ Ke7 47.bxa5 Qc3+ 48.Kh4 Qc2 49.Qe5

Jiimfromprovidence, computer agrees that 29 Nd6 is best, but it agrees with Honza that black is still badly lost.